using hyperspectral imaging in nuclear radiation …
TRANSCRIPT
644
Grégory Lucas1 – Solymosi József2 – Lénart Csaba3
USING HYPERSPECTRAL IMAGING IN NUCLEAR RADIATION
AERIAL RECONNAISSANCE? A PRELIMINARY STUDY45
This article aims at exploring the potential of hyperspectral imaging in order to develop further the Hungarian
aerial nuclear reconnaissance system. First a description about the theoretical basis of ionizing radiations is
provided. The different types of radiations, their characteristics, penetration range and effect on matter are
presented. Then a critical analysis is done on the Hungarian nuclear reconnaissance system based on the detection
capacities and operational implementation criteria. The last part explores the potential of hyperspectral imaging
technology, its added values and the different possibilities envisaged for its application.
HIPERSPEKTRÁLIS KÉPALKOTÁS ALKALMAZÁSA A NUKLEÁRIS SUGÁRZÁS LÉGI
FELDERÍTÉSÉBEN? ELŐZETES TANULMÁNY
A cikk célja a hiperspektrális képfelvételezés potenciáljának felmérése a magyar légi nukleáris felderítő rendszer
továbbfejlesztése érdekében. Elsőként az ionizáló sugárzás elméleti alapjairól adunk leírást. Bemutatjuk a
sugárzás különböző típusait, azok jellemzőit, behatolási tartományaikat, és az anyagra gyakorolt hatásaikat.
Ezután egy kritikai elemzést végzünk a magyar nukleáris felderítő rendszerről, a felderítési kapacitásokra
alapozva. Az utolsó rész a hiperspektrális képalkotási technológia lehetőségeit deríti fel, annak hozzáadott értékeit
és megvizsgálandó alkalmazási lehetőségeit.
1. INTRODUCTION
Aerial nuclear reconnaissance is a survey technique used to detect, measure, identify and map
radioactivity in the environment. The main field of application is the estimation of the
contamination extent after a nuclear catastrophe or nuclear attack. Another possible application is
the localization of lost punctual radioactive sources. All the aerial nuclear reconnaissance
techniques are based on gamma radiation detection because it is the most penetrating radiation type.
In average, in order to detect moderately intense radiations, the reconnaissance is done at an altitude
of 200m above the ground. This is because of the quick attenuation of the gamma radiation by the
air. In this study we would like to explore the opportunity offered by an optical remote sensing
technique called hyperspectral imaging. Our assumption is that with optical remote sensing
technique it could be possible to detect a larger range of radiations from a higher flight altitude.
1 National University of Public Service, Doctoral School of Military Engineering, [email protected] 2 Dr., Karoly Robert College, Institute of Remote Sensing and Rural Developement, [email protected]
3 Col(Ret.) Prof. Em., National University of Public Service, Institute for Disaster Management,
[email protected] 4 Publisher’s reader: Laszlo Pokoradi (PhD), Professor, University of Debrecen, [email protected] 5 Publisher’s reader: Colonel Robert Szabolcsi (PhD), Professor, National University of Public Service
Technology Department of Military Aviation, [email protected]
645
2. CHARACTERIZATION OF IONIZING RADIATIONS
Ionizing radiation is radiation composed of particles that individually carry enough energy to
liberate an electron from an atom or molecule. This involves the ejection of an orbital electron,
resulting in the creation of an ion pair.
X → X+ + e– (generic formula for single ionization)
Ionizing radiation includes cosmic rays, alpha, beta and gamma rays, X-rays, and in general
any charged particle moving at relativistic speeds. In the present study we consider ionizing
radiations generated through nuclear reactions, either artificial or natural. [1]
2.1. The different types of ionizing radiations
Radiations can be grouped into directly ionizing radiations and indirectly ionizing radiations
(Tab.1). Directly ionizing radiations include all charged particles such as alpha particles, beta
particles and heavier ions. All charged particle radiations lose energy interaction with the orbital
electrons or nuclei of atoms in the materials they traverse. Indirectly ionizing radiations include
some types of electromagnetic radiations and neutrons. These radiations interact with matter by
giving rise to secondary radiation which is ionizing. Indirectly ionizing radiations lose energy
by collisions with electrons, or atomic nuclei, and the charged particles thus set in motion
interact in turn with the orbital electrons or nuclei. [1]
Type of radiation Ionizing radiation Elementary charge Mass (MeV/c2) Electromagnetic
radiation
Indirectly ionizing ultraviolet 0 0
X ray
Gamma Ray
Particles Neutron 0 940
Directly ionizing Electron / particle β- -1 0,511
Positon / particle β+ +1 0,511
Muon -1 106
Proton +1 938
Ion 4He / particle α +2 3730
Ion 12C +6 11193
Other ions Variable Variable
Tab. 1. Main ionizing radiations and their characteristics
2.2. Penetration ranges and consequences for airborne detection
The aerial detection of ionizing radiations is governed by two main principles which are the
interaction of the ionizing radiations with matter (both with the sensor and the air) and the
penetration of the ionizing radiation. [1]
Charged particles such as electrons, positrons, protons, alpha particles and beta particles
strongly interact with electrons of an atom or molecule. Consequently they have a very low
penetration range in matter. For example Alpha particles are absorbed by about 10-2 m of air
and the penetration range of Beta particles is about 8 m in air. [1] Those radiations stopped far
before the sensor cannot “activate” crystals or induce ions pair production in the chamber of
scintillation or semi-conductor sensor used in classical airborne applications. With the present
technology and practices those radiations are lost for the detection process.
646
Neutral particles like gamma and neutron interact less with matter, are indirectly ionizing and
have the higher penetration range. The penetration range of gamma radiation is several
hundreds of meters in the air. The penetration of neutron radiation depends on the content of
water of the air, as water shields neutron radiation. Neutron radiation can induce gamma
radiation after atomic activation. [1]
2.3. Effect on matter and detection strategies
Let’s first confront the two different detection strategies considered in this study: Airborne
Gamma Spectroscopy (AGS) and remote sensing optical methods.
AGS has been recognized as a very powerful tool for the detection of ground contamination and
to locate lost radioactive sources. It is presently the method recommended by the IAEA for the
detection of ionizing radiations. [2] Optical detection of radioactivity by remote sensing is still
under experiment and has not proved yet efficiency in ionization detection. [6][7]
The two methods lay on totally different measurement principles. AGS detection principle is based
on the “capture” of energetic photons by the detector, which means all the photons from gamma
radiation stopped in the air prior the sensor are lost for the detection process and the low penetration
range radiations (alpha and beta) are lost too. Individual radionuclide emits gamma rays of specific
energies that are characteristic for an element and isotope. Gamma ray measurements can be
conducted in two modes. Total count measurements register gamma rays of all energies. These are
used to monitor the gross level of the gamma radiation field and to detect the presence of anomalous
sources. Spectrometers, on the other hand, measure both the intensity and energy of radiation (Fig.
1), and this enables the source of the radiation to be diagnosed. Gamma ray spectrometry is thus a
powerful tool for monitoring the radiation environment. [2]
Fig. 1. Example of spectrogram generated by gamma spectrometry
Optical remote sensing methods can in theory detect what happens in the air and in the vicinity
of the radioactive source (which encompass the effect of alpha and beta radiations on matter) as
the detection medium (the electromagnetic radiations) is traveling almost freely in the air. The
detection principles should allow detecting electronic excitation generated by the ionizing
radiations (emission) and the molecular species generated by the ionization of the air (through the
647
radiations absorbed, reflected or transmitted and the associated spectral signature).
In conclusion the two detection methods are complementary and could be used in combination
to improve the detection range and the accuracy of the detection.
3. CRITICAL DESCRIPTION OF THE HUNGARIAN NUCLEAR
RECONNAISSANCE SYSTEM
The gamma spectrometry airborne nuclear reconnaissance system was designed for the primary
survey of area contaminated by radiological materials. The system can be used to reach three
major goals:
to measure the extended contamination on a territory and to map it. This is made with
total count measure of radiation rate.
the localization of radioactive point sources.
the identification of radioactive isotopes. [3][4][5]
3.1. System built up
The container of the system includes two nuclear detectors, a GPS-receiver, a barometric
altitude-meter and data logger which can send the recorded data to the on-board notebook or to
the PC of the operation center. One of the two detectors is a Geiger Müller tube (BNS-98) dose
rate meter while the other one is a specially designed, highly sensitive NDI-65/SK type
intelligent scintillation detector, built in a lead collimator which ensures the capability of
finding and localizing discrete radiation sources on the ground. The system calculates the
radiation level of the contaminated area (referred to 1 m altitude) or the dose rate of a discrete
radiation source from the dose rate measured at the flying altitude considering the atmospheric
and ground conditions.
Fig. 2. Details about the composition of the nuclear reconnaissance system
648
3.2. Characteristics of the system
Name Sensors in the system Operating mode Detection capacities
LABV
airborne
-BNS-98 (2s) remote
dose-rate meter (GM
tube);
- NaI(Tl) (0,5s) crystal
NDI-65/SK intelligent
scintillator in plumb
collimator
Helicopter (MI-24D)
uniform contamination:
Speed: 150-180 km/h
Altitude: 80-100 m.
Coverage: 300 km2/h
Point source:
Speed: 100-120 km/h
Altitude: 50-60 m.
Coverage: 18-20 km2/h
Point sources:
1,5-2 times the natural ground
value 2-3 times the natural ground
value for radiation level
uniform activities:
over 2-5 mGy/h of dose rate
over 10-20 μGy/h count rate
Tab. 2. Main characteristics of the Hungarian reconnaissance system.
The main limitations identified is the low flying altitude required to measure gamma radiation, from
50m to 200m above ground. Another limitation is the fact that low penetration radiations (which are
stopped by the air in the vicinity of the radioactive materials) cannot be detected and are lost for the
detection process. It should also be noticed that the reconnaissance system is not performing geo-
referenced imaging of the impacted area during the flight, which in case of catastrophe management
could be a source of relevant information for evaluating the impacts on environment and population.
4. PRESENTATION OF HYPER SPECTRAL IMAGING TECHNOLOGY,
WORKING PRINCIPLE AND AVAILABLE SENSOR
Hyper spectral images are produced by instruments called imaging spectrometers. The
spectrometers measure the energy received simultaneously in hundreds of narrow (several nm)),
adjacent spectral bands. These measurements make it possible to derive a continuous spectrum
for each image cell. Spectroscopy science analyzes how reflectance varies with wavelength in
a spectrum. A spectrum is like a fingerprint where are appearing spectral domains of low and
high reflectance as a consequence of the physico-chemical properties of the materials surveyed.
By the identification of characteristic absorption or reflection patterns it is possible to determine
which materials are imaged. [11]
Fig. 3. General principle of hyperspectral imaging - ©MicroImages, Inc., 1999-2012
649
Image spectra can be compared with field or laboratory reflectance spectra in order to recognize
and map surface materials such as particular types of vegetation or diagnostic minerals
associated with ore deposits. Wavelength-specific absorption may also be caused by the
presence of particular chemical elements or ions or the ionic charge of certain elements.
Reflectance varies with wavelength for most materials because energy at certain wavelengths
is scattered or absorbed to different degrees. [11]
Once should distinguish the reflected light spectroscopy method where the spectral reflectance
(ratio of reflected energy to incident energy as a function of wavelength) is measured and the
emission spectroscopy where the electromagnetic emission from elements or chemical species
are measured.
The most important characteristics of imaging spectrometers are:
spectral range
spectral resolution
spatial resolution
signal-to-noise ratio
The Tab. 3. shows a summary of the airborne sensors owned by Karoly Robert College.
Sensor name Sensor type Spectral range
AISA Eagle Hyperspectral sensor (VNIR) 400-970 nm
AISA Hawk Hyperspectral sensor (SWIR) 970-2450 nm
Tab. 3. Available sensors and their detection ranges
The AISA Eagle and Hawk respectively cover the 400-970 nm and 970-2450 nm spectral range.
Their technical characteristics differ regarding the spectral resolution: 2,9 nm for Eagle and 8,5
nm for Hawk. When used at the highest spectral resolution the dual sensor collects 498 bands
in the 400-2450 nm region. Spectral binning which consist in regrouping spectral bands is
possible with the two sensors and offer a stronger signal if there is a strong response in one part
of the spectrum. The detailed specifications are provided in Tab. 4.
Tab.4. Technical specifications of the AISA dual hyperspectral sensor by SPECIM
Natural color images and RGB-IR images (orthophotos) can be derived from hyper spectral
650
imaging. This additional source of information is of high relevance for catastrophe
management.
Hyperspectral imaging had demonstrated many applications in resource management,
agriculture, mineral exploration, monitoring of vegetation and contamination detection. At
present no application was attempted for the detection of ionizing radiations but some
successful applications at the margin of our topic could be adapted to fit our specific objectives.
For example some results were published on successful LWIR identification of hazardous
gasses. [16] LWIR hyperspectral imaging is also capable in identifying chemicals used in
chemical warfare (Farley et al, 2006). [17] FTIR imaging has been successfully used to identify
radioactive materials. [8] Some preliminary works has been done with hyperspectral remote
sensing for the identification of uranium mine tailings.
The work recently done for the detection of soil contamination with the red mud catastrophe in
Kolontar also shows some interesting potential with the detection of contaminant in low
concentration in soil. [13]
Last but not least, hyperspectral technology is evolving very quickly. The sensors developed
recently have a made a significant progress with signal-to-noise ratio and spectral resolution.
This open new possibilities for the detection of traces of gas and molecules. [9][14]
5. STUDY ON THEORETICAL PHYSICAL BASIS
In the previous chapter we have described the general principles regarding hyperspectral
technology and have introduced one sensor as an example. In the light of the additional
explanations given about the ionizing radiations and their interaction with matter we are trying
in this chapter to set some basis for the indirect detection of ionizing radiations with
hyperspectral imaging technology.
5.1. The strategical basis
As we have seen previously, presently the aerial detection of ionizing radiation is only done with
Aerial Gamma Spectrometry (AGS). This method specifically senses the high energy photons
generated by the decay of radiological materials (only gamma unscattered radiation). Only
photons have a sufficient penetration range to travel in the air and reach an airborne sensor. Alpha
and Beta radiations which are respectively stopped by a few cm and 9 cm of air are lost for such
detection process. Gamma radiation intensity decreases exponentially with altitude. Because of
this reason, the sensor should be flown at an average altitude of 100-200m (with helicopter) for
moderately intense radiation sources, which is a main disadvantage: it is costly and lack of
flexibility as regards to the new challenges in nuclear reconnaissance.
We would like to develop a new detection method based on a different strategy. Instead of using
gamma radiation detection, we would like to use an optical remote sensing approach and to detect
the ions and molecules specifically generated in the air by the ionization radiations around
radioactive materials. This approach would be done with reflected light spectrometry. This is an
indirect measurement method, but it offers two advantages. First the effects of alpha, beta and
651
scattered gamma radiations on matter would be sensed. The activity of radioactive materials could
then be retrieved from appropriate calibration and computation. As a consequence hyperspectral
imaging could be used as a complementary method. AGS and hyperspectral imaging would detect
(indirectly for the sake of hyperspectral imaging) the full range of radiations in the vicinity of the
radiological material and in the air from the radiological source to the sensor. Secondly, as light
travel more freely in the air, the flying altitude could be increased. Puckrin have demonstrated
that in the case of passive detection with FTIR, radiation can in theory be detected from an altitude
of 1000m above the ground if the conditions are optimal. [8] The demonstration was made using
MODTRAN4 modeling. [8]
Fig. 4. and Tab. 5. emphasizes on the difference in the application of Airborne Gamma
Spectrometry and the application of hyperspectral imaging as regard to the flight altitude, the
platform used, the “objects” sensed, the medium used. It should be noticed that all the
information related to the implementation of hyperspectral imaging are only hypothetical as
this method was not put in practice yet.
Fig. 4. Comparison of Airborne Gamma Spectrometry and hyperpectral imaging for the detection of ionizing
radiations.
652
Airborne Gamma Spectrometry Airborne hyper spectral imaging Detected Not detected and lost for
detection process
Potentially detected Not detected and
lost for detection
process
Vector Photons (gamma) Molecules and excited
atoms
Radiation Gamma ray from:
- unscattered gamma
- neutron activation.
Alpha, beta, scattered
gamma, neutron (partly)
and charged particles.
Partly: alpha, beta,
scattered gamma,
neutron, charged
particles.
Unscattered
Gamma (photons
of high energy not
stopped by the air).
Effect on
matter
X+,e- pairs creation on
the crystal of the
sensor.
X+,e- pairs created in the air
before the sensor.
Atomic excitation.
Atomic excitation.
Specific species created by
ionizing radiation (ions,
secondary product (O3))
-
Detection
range
On the crystals of the
sensors. Only the
photons intercepted by
the sensors are counted.
On the surface of the
radioactive material, in the
vicinity of RA material, in
the air between the ground
and the sensor.
In the field of view of
the sensor, i.e. in the
vicinity of radioactive
material in the air or on
the ground.
Over the sensors
(photons with high
energy)
Physical
effect used
in the
detection
Ionization and
scintillation created by
photons by:
- photoelectric effect,
- Compton scattering,
- pair production.
Electromagnetic
radiation emission (after
de-excitation of atomic
electrons), absorption
and reflexion (by the
product of ionization
reactions).
Tab. 5. Comparison of Airborne Gamma Spectrometry and hyperpectral imaging for the detection of ionizing
radiations.
The strategy with the use of hyperspectral technology can be twofold:
to detect the presence of products generated by ionization (molecules species) by
reflected light spectrometry. The absorption and reflection patterns of the products
specific to ionization reactions should then be known and identified.
to exploit the excitation generated by the ionizing radiations, which means to measure
the energy emitted by excited atoms when returning to ground state (O, N, H).
5.2. Detection of new molecular species by reflected light spectroscopy
From the bibliographic research we know that Moss already attempted to detected ionizing
radiations through the detection of new species generated by radiation with optical remote
sensing method. [6] If the strategy is the same as the one we want to develop, the method differs
as he used differential absorption LiDAR (DIAL) for the detection. Nevertheless Moss made
an interesting exploration regarding the specific species present in the surrounding of
radioactive materials. He has calculated the molar fraction as a function of time of the species
formed in air by irradiation with a 60 Curie source of 113Cd. The simulation was done with a
gas-phase chemical kinetics code developed at the Los Alamos National Laboratory. The results
of the simulation are very interesting. Instead of having ions as a product of the ionization of
air (which is expected primarily from ionization), the model shows that rather secondary
products are accumulated in the vicinity of the radiological materials. The model generated the
following secondary products: O3, N2O, HNO3, H2NO2, and NO2. The second important result
is the value of the molar fraction calculated by the model. They are very low, from the order of
10-5 to 10-7. The absence of ions is probably explained by their high reactivity and very quick
653
life time in the air. Once the specific indicator species are theoretically known, we should
explore the detection possibilities. In this article we decided to concentrate on O3 and N2.
The classical absorption bands of ozone are the following:
the Hartley bands between 200 and 300 nanometers in the ultraviolet, with a very intense
maximum absorption at 255 nanometers. It is the strongest absorption band.
the Huggins bands, weak absorption between 320 and 360 nanometers
the Chappuis bands, a weak diffuse system between 375 and 650 nanometers in the
visible spectrum
the Wulf bands in the infrared beyond 700 nm, centered at 4,700, 9,600 and 14,100
nanometers, the latter being the most intense.
The figure bellow represents the absorption bands for ozone and the different domain of
hyperspectral sensors.
Fig. 5. Match between the absorption of ozone and the light spectrum sub-regions.
Some absorption is located in the VNIR and MWIR regions. The highest potential is in the
LWIR region comprising the 14100 nm band (Fig. 5.).
Regarding the other target species, additional research about their spectral signature should be
done in order to know if it makes sense to try to detect them.
5.3. Electronic excitation
Regarding nitrogen, the following reactions happen:
N2→ N2* N2* → N2 + hv
References are available about the atomic emission line from spectroscopy analysis. Atomic
emission spectroscopy is a method of chemical analysis used in laboratory for identifying the
elements in a sample. The principle reposes on the emission of photons by excited atoms.
Fig. 8. and 9. represent the atomic emission spectrum of Nitrogen. They were elaborated from
the atomic basic spectroscopic data provided by the National Institute of Standards and
Technology (NIST). [15] Only the observed strong emission lines are represented on the figure.
654
Fig. 6. Atomic emission spectrum for Nitrogen I-II in the spectral range of Eagle (4000-9500 Å).
Several spectral regions show a high density of emission lines. This is of interest for the
detection with the hyperspectral sensors. In the spectral range of Eagle the 460, 500, 570, 750,
820, 870 and 940 nm regions seem promising.
Fig. 7. Atomic emission spectrum for Nitrogen I-II in the 9500-14 000Å region.
The emission spectrum of nitrogen in the spectral range of Hawk seems quantitatively less
important. Emission lines in the 1010 and 1350 nm region could offer some detection
possibilities.
The excited state of oxygen is somewhat more stable than nitrogen. While de-excitation can
occur by emission of photons, more probable mechanism at atmospheric pressure is a chemical
reaction with other oxygen molecules, forming ozone.
O2 → O2+ + e–
O2+ + 2O2 → 2O3
The detection possibilities are then the same as the ones already exposed above. The creation
of ozone molecule by both excitation and ionization is an interesting fact for the detection
capacities as it can strengthen the absorption signal.
655
5.4. Limitations and problem expected with hyperspectral imaging
As mentioned, only traces of molecular species are expected in the air around the radiological
materials. We wonder about the intensity of the spectral signature the sensor can detect from
the specific species. The question of the relative intensity compared to the other spectral
signatures (background) is a key point for the extraction of the spectrum of the specific species.
If the reflectance is too weak, the atmosphere could also create too much disturbances and their
spectrum could not be extractable. A last question is the spectral accuracy necessary to be sure
to see the spectral signature. Is the scale of 3-4 nm sufficient or should the sensor have a sub-
nanometer spectral resolution? Laboratory and field test with different sensors will try to answer
these questions.
6. CONCLUSIONS AND PERSPECTIVES
The theoretical analysis conducted on hyperspectral technology reveals an interesting potential
for the detection of alpha, beta and scattered gamma radiations through the detection of specific
signal signatures from the generated ions and secondary products of ionization reactions.
Consequently hyperspectral imaging potentially constitute a complementary detection method
to aerial gamma spectrometry. The integration of the two detection methods on the same
platform would allow an integrated approach in the detection of ionizing radiations.
Furthermore this optical remote sensing technique can be applied at an altitude much higher
than 100m. Nevertheless several difficulties have been identified. It seems the products of
ionization are present in the air in infinitesimal quantities. Laboratory and field measurements
work should confirm if optical detection is sensitive enough and applicable. The strength of the
radiological source and the distance with the source are important parameters to consider in this
work. A second question deals with the reflectance and spectral signature of ionized molecules
and secondary products, in particular the intensity and profile of the signals. A weak or “flat”
signal would offer limited applications as the spectral signature could not be extracted from a
mixed spectrum comprising the general environmental effects (atmospheric and background).
The laboratory work in the near future should help to determine which spectral region is the
most promising and which indicator species have the best potential for ionizing radiation
detection.
Hyper spectral imaging offers two other added values. Orthorectified RGB-IR images can be
produced and used for estimating the damage on the environment. The spectral signature of
vegetation in the infrared red region can be used as an indicator of stress and help in the
identification of radiological contamination.
656
REFERENCES
[1] V. VALKOVIĆ: Chapter 5 - Measurements of Radioactivity. Radioactivity in the Environment. Amsterdam,
Elsevier Science: 117-258, 2000.
[2] IAEA: Guidelines for radioelement mapping using gamma ray spectrometry data, 2003
[3] J. SOLYMOSI, E. BAUMLER, A. SARKADI, Á. GUJGICZER, I. PINTÉR, Á. VINCZE: Wide range universal
radiation measuring instrument. AARMS, Volume 1(1) (2002) 133–144.
[4] J. ZELENÁK, J. CSURGAI: Analysis of the applicability of the airborne radiological reconnaissance in case of
searching lost or stolen radioactive sources. Hadmérnök, (2009) 46-62.
[5] J. SOLYMOSI, E. BAUMLER: Eljárás és berendezés ismeretlen összetetélű és/vagy többkomponensű, főként
hasadási termékekkel kontaminált terepszakaszok sugárszintjének légi felderitésére, Hungarian Patent
201161 B, issued 1990-09-28.
[6] C. E., MOSS, R. M. GOELLER, D. F. MILLIGAN, J. E. VALENCIA, J. ZINN: Remote sensing of radiation.
Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors
and Associated Equipment, 422(1-3) (1999) 832 – 836.
[7] S. IHANTOLA, J. SAND, K. PERAJARVI, J. TOIVONEN, H. TOIVONEN: Fluorescence-Assisted Gamma
Spectrometry for Surface Contamination Analysis. IEEE Transactions on Nuclear Science,,
Vol. 60 , Issue: 1 , Part: 2 (2013).
[8] E. PUCKRIN, J. THÉRIAULT: Passive standoff detection of radiological products by Fourier-transform
infrared radiometry, Opt. Lett., 29 (2004) 1375-1377.
[9] H. HOLMA, A. J. MATTILA, & T. HYVÄRINEN: New thermal infrared hyperspectral imagers. NATO
Technical report, RTO-SET-151, 2009.
[10] J. TUOMINEN, T. LIPPING: Detection of Environmental Change Using Hyperspectral Remote Sensing at
Olkiluoto Repository Site, 2011.
[11] R. B. SMITH: Introduction to hyperspectral imaging, 2012.
[12] J. LÉVESQUE, R. A. NEVILLE, K. STAENZ, Q.S. TRUONG: Preliminary results on the investigation of
hyperspectral remote sensing for the identification of uranium mine tailings. In: Proceedings of the ISSSR:
June 10-15, 2001, Quebec City, Canada
[13] Cs. LENART, P. BURAI, A. SMAILBEGOVIC, T. BIRO, Zs. KATONA, R. ANDRICEVIC: Multi-sensor integration
and mapping strategies for the detection and remediation of the red mud spill in Kolontar, Hungary:
Estimating the thickness of the spill layer using hyperspectral imaging and Lidar, Hyperspectral Image and
Signal Processing. Evolution in Remote Sensing (WHISPERS). 3rd Workshop, Lisbon, Portugal. (2011).
[14] R. RICHTER: Hyperspectral Sensors for Military Applications. Emerging EO Phenomenology, 2005.
[15] J. E. SANSONETTIA, W. C. MARTIN: Handbook of Basic Atomic Spectroscopic Data. J. Phys. Chem, Ref.
Data Vol. 34(No. 4) (2005)1885-1890.
[16] G.M. GITTINS, W. J. MARINELLI: LWIR multispectral imaging chemical sensor. In proc. SPIE 3533, Air
Monitoring and Detection of Chemical and Biological Agents, 93, Boston. (1999).
[17] V. Farley, A. Valleres, M. Chamberland, A. Villemaire: Performance of the FIRST, a longwave infrared
hyperspectral imaging sensor. Proceedings of SPIE optically based biological and chemical detection for
defense, Stockholm, Sweden. (2006).